• Resolved

AFE4300EVM-PDK: Why does body impedance decrease over time when measuring in BCM mode?

Part Number: AFE4300EVM-PDK

I think you should have seen a lot of these problems because this chip is a chip for measuring human impedance.
Hi

For BCM measurement, I set it as below picture(PICTURE1, PICTURE2)

and To measure human impedance, i connected the AFE4300EVM-PDK and the electrodes as shown below(PICTURE3, PICTURE4).


you can see the yellow cable connection.


As i measure up over the electrode, the voltage value corresponding to the impedance continues to decrease with time!!!  (You can see the PICTURE5)


After calibration, the impedance difference between the initial voltage value and the last voltage value corresponds to about 50 ohms. it is impossible.


When connecting to a passive element(Resistor,Capacitor), the voltage value is constant over time.


Only when a person is up, the voltage continues to decrease with time.


I work for companies that make medical devices that use impedance.

I would like to use the AF4300 for measuring human impedance, but I can not solve this problem.


I think you should have seen a lot of these problems because this chip is a chip for measuring human impedance.


May I ask for your advice?

Picture. 1

picture.2

picture 3

picture 4

picture 5

  • Hi Sehwan,

    I see you have two thread for same topic, can you please delete/ close the other one?

    Regarding you problem, we have not seen such behavior. Following information could provide some clue.

    1) It might be due to variation in electrode-skin contact impedance. Is it same phenomenon across people?
    2) If you wait sufficiently long, does the value becomes stable?
    3) What is the BCM_DAC_FREQ? Is it FWR mode of IQ mode. Can you please try for different frequencies as well as different mode.

    Regards,
    Prabin
  • Sorry for the late reply.

     

    1) It might be due to variation in electrode-skin contact impedance. Is it same phenomenon across people?
    A: Other people still have impedance decreasing over time.

     

     

    2) If you wait sufficiently long, does the value becomes stable?

    A: I stood for 17minutes like the picture 5, but it did not converge.


    3) What is the BCM_DAC_FREQ? Is it FWR mode of IQ mode. Can you please try for different frequencies as well as different mode.

    A: As you can see in the picture.1, the frequency is 20kHz

     

    And we made human impedance model like picture 6

    According to the formula, the impedance value at 20 kHz should be 538 ohms. Like picture7

    However, after connecting the measurement as shown in the picture 8, the result obtained after calibrating the result obtained from 300 ohms, 500 ohms, and 700 ohms results in an impedance of 595.1 ohms. The calculation process is shown in the picture 9

    picture6

    picture 7

    picture 8

    picture 9

    The same is true for 100kHz

    According to the formula, the impedance value is 472 ohms.

    the result obtained after calibrating the result obtained from 300 ohms, 500 ohms, and 700 ohms results in an impedance of 488 ohms.

     

     

     

    Perhaps the larger the frequency, the smaller the reactance of the capacitor, which reduces the difference between the formula value and the calibrated result, but still has a large error.

     

    Our company actually measures the impedance value through the calibration. The error is too large to calculate the impedance as it is.

    Although the error is large, the rate of decay over time when measuring impedance through a human resistance model is much smaller than when measuring people.

    And I attached the files of 20kHz, 100kHz and IQmode result screen shot files as files.AFE Result.zip

  • In reply to sehwan park:

    Hi Sehwan,

    Thank your for the detail explanation.
    However i have following concern.
    1) Are the values of components in the human body model exactly same as mentioned? What is the tolerance level? Will you be able to measure them with external LCR meter.
    2) Just to remove any mismatch ( due to parasitic caps because of wires) between the calibration path and body impedance path, can you put the human body model on the EVM and try.

    Regards,
    Prabin